Are there ICs specifically for noise reduction?

Thread Starter

Green Bean

Joined Mar 31, 2017
126
So, if you had a jittery signal that you wanted to be constant and noise-free, are there ICs that you could attach to reduce the noise before the signal is outputted to a device?
 

Papabravo

Joined Feb 24, 2006
21,159
Couple of questions:
  1. What kind of "jittery signal" are you talking about?
  2. What do you mean by "constant"?
  3. What kind of device is the signal going to?
In general noise can be reduced or attenuated, but it cannot be eliminated -- sorry.
 

nsaspook

Joined Aug 27, 2009
13,087
If you have clocked digital signals they can be Regenerated back to the exact original if the signal falls within the error and S/N requirements of the regeneration device. We used them frequently to remove HF transmission effects (frequency selective fading or multipath) on encrypted teletype signals before re-transmission on another comm link.

Pulses can be re-timed to remove jitter (arriving too soon or too late), statistical techniques can be used to re-construct data from a noisy signal but sometimes all you get is GIGO.

What circuit or IC might be used depends on the type and complexity of the 'noise' component and signal. Sometimes a simple Schmitt Trigger circuit is all you need to restore the proper signal.
 

Thread Starter

Green Bean

Joined Mar 31, 2017
126
Couple of questions:
  1. What kind of "jittery signal" are you talking about?
  2. What do you mean by "constant"?
  3. What kind of device is the signal going to?
In general noise can be reduced or attenuated, but it cannot be eliminated -- sorry.
1. Jittery as in, you expect 5 volts but get random voltages between like 4.8 and 5.2 or something. Ya know, noisy.
2. Constant as in, you expect 5 volts and get 5 volts.
3. In my case, a speaker, but it could be anything.

I want to send an oscillating signal out the joystick port of my Commodore 64, by writing 0 and 1 to the bits in the register that corresponds with the port. The problem is that the port is a little noisy, so it doesn't sound that great.
 

DickCappels

Joined Aug 21, 2008
10,152
I would try sending the signal through a buffer (7404, 7414, or something like that) and heavily filter the power supply to that chip to remove the noise.

Edit:
Alternatively you c;an uses a simple clipper thought its output would need to be amplified in order to drive a speaker.

upload_2017-5-28_8-36-5.png

There are countless variations of clippers and limiter circuits on the internet.
 
Last edited:

Papabravo

Joined Feb 24, 2006
21,159
1. Jittery as in, you expect 5 volts but get random voltages between like 4.8 and 5.2 or something. Ya know, noisy.
2. Constant as in, you expect 5 volts and get 5 volts.
3. In my case, a speaker, but it could be anything.

I want to send an oscillating signal out the joystick port of my Commodore 64, by writing 0 and 1 to the bits in the register that corresponds with the port. The problem is that the port is a little noisy, so it doesn't sound that great.
The reason I asked is that "jitter" usually refers to changes in phase or frequency of a signal. If you are talking about a digital signal going to a speaker it is going to sound like crap regardless of what you do. Have you actually observed the "jitter" you spoke of on an oscilloscope? 400 mV peak to peak seems unusual for a signal that normally goes between +5V and GND. Using a CMOS buffer may help you, but haphazard layout and grounding just might introduce additional problems. In my experience digital signals are hardly ever constant and in most case it doesn't matter since only the logic thresholds are important for digital purposes. It might be possible to use some RC filtering to round off the square edges of your digital signal going to the speaker, but that is about the best you can hope for.
 

Thread Starter

Green Bean

Joined Mar 31, 2017
126
The reason I asked is that "jitter" usually refers to changes in phase or frequency of a signal. If you are talking about a digital signal going to a speaker it is going to sound like crap regardless of what you do. Have you actually observed the "jitter" you spoke of on an oscilloscope? 400 mV peak to peak seems unusual for a signal that normally goes between +5V and GND. Using a CMOS buffer may help you, but haphazard layout and grounding just might introduce additional problems. In my experience digital signals are hardly ever constant and in most case it doesn't matter since only the logic thresholds are important for digital purposes. It might be possible to use some RC filtering to round off the square edges of your digital signal going to the speaker, but that is about the best you can hope for.
I see. I plan on getting a DAC so that I can experiment with different waveforms (aside from square waves). I was really just messing around with this pure digital thing, my end goal is to make a programmable waveform generator. Now that I think of it though, wouldn't a DAC get rid of some of the noise? If I send a somewhat noisy digital signal in, it'd still produce a constant voltage, right?

IMG_20170527_2203316_rewind.jpg

It's a crude illustration, but you can see that even though the digital "1s" aren't always represented with exactly 5v (because of noise), it's close enough that you still end up with a stable output. So by "filtering" through the DAC, I could probably significantly reduce the noise and get a much cleaner square wave (and be able to make other waveforms as well).
 

nsaspook

Joined Aug 27, 2009
13,087
I see. I plan on getting a DAC so that I can experiment with different waveforms (aside from square waves). I was really just messing around with this pure digital thing, my end goal is to make a programmable waveform generator. Now that I think of it though, wouldn't a DAC get rid of some of the noise? If I send a somewhat noisy digital signal in, it'd still produce a constant voltage, right?

View attachment 127715

It's a crude illustration, but you can see that even though the digital "1s" aren't always represented with exactly 5v (because of noise), it's close enough that you still end up with a stable output. So by "filtering" through the DAC, I could probably significantly reduce the noise and get a much cleaner square wave (and be able to make other waveforms as well).
Most likely a commerical DAC chip would not use the input digital signals to directly derive the analog output levels.
A old AD7521 example: http://www.mouser.com/catalog/specsheets/intersil_fn3104.pdf
http://www.analog.com/media/en/training-seminars/tutorials/MT-015.pdf

You can experiment with a resistor ladder R-2R DAC.
http://www.tek.com/blog/tutorial-digital-analog-conversion-–-r-2r-dac
http://www.mouser.com/Passive-Components/Resistors/Resistor-Networks-Arrays/_/N-e89l?P=1z0sqtq
 

Thread Starter

Green Bean

Joined Mar 31, 2017
126
So it looks like the voltages of the inputs would directly effect the output of a resistor ladder. But, instead of sending the digital signal directly to the DAC, I thought I could send it to the bases of some transistors. The collectors would go to VCC, and the emitters to the DAC. Since VCC is constant (no noise) this would filter the noise out right? Because the noisy signal is just used to activate the transistors (being used as switches) not actually to create the analog output.

Does this sound ok?
 

Papabravo

Joined Feb 24, 2006
21,159
So it looks like the voltages of the inputs would directly effect the output of a resistor ladder. But, instead of sending the digital signal directly to the DAC, I thought I could send it to the bases of some transistors. The collectors would go to VCC, and the emitters to the DAC. Since VCC is constant (no noise) this would filter the noise out right? Because the noisy signal is just used to activate the transistors (being used as switches) not actually to create the analog output.

Does this sound ok?
Power supplies can be noisy. Especially if there are load transients which will happen when digital logic IC's are powered by the power supply. You need to get an oscilloscope and start poking around in circuits so you have some idea of what to expect.
 

joeyd999

Joined Jun 6, 2011
5,237
So it looks like the voltages of the inputs would directly effect the output of a resistor ladder. But, instead of sending the digital signal directly to the DAC, I thought I could send it to the bases of some transistors. The collectors would go to VCC, and the emitters to the DAC. Since VCC is constant (no noise) this would filter the noise out right? Because the noisy signal is just used to activate the transistors (being used as switches) not actually to create the analog output.

Does this sound ok?
Typically, a DAC uses a low-noise stable voltage reference from which to derive it's output.

How many bits do you require, and how much noise are you willing to deal with?
 

Thread Starter

Green Bean

Joined Mar 31, 2017
126
Power supplies can be noisy. Especially if there are load transients which will happen when digital logic IC's are powered by the power supply. You need to get an oscilloscope and start poking around in circuits so you have some idea of what to expect.
I'm only 15 so I don't have a job. I have $40 to my name lol. What's the cheapest scope I can get?

I've been using an app on my tablet but I don't think it's very accurate.
 
Top